Innovation starts with knowing your customers well. As the number of interaction channels with customers’ increases, the challenge of aggregating information and insights from such customer interaction sources becomes increasingly complex.
Banks are relying on various channels for end-user touch points. Today banks use a mix of traditional and contemporary channels that provide customers an ability to enquire, complain or request for services. The traditional channels being email, websites and customer (call) centers. With today’s trend being social, banks are also embracing social media as a means for customers to reach out to them. Social media provides the dual advantage of being a quick information dissemination tool for the banks. Although banks have invested in setting up infrastructure for these touch points, use of information generated from these touch points is still not optimized. Lack of cross platform information sharing, coupled with silo approach to relationship ownership creates barriers for having a 360° degree view of the customer, thereby losing out of tremendous knowledge and insights.
Customer touch points are notorious for generating unstructured data. Due to the nature of this data, often companies struggle to link data across various channels. This is where innovation is required. Top companies have started investing in unstructured data analysis, to link customer activities with their interests and satisfaction with the system.
It is important to understand and analyze unstructured data from disparate sources as a whole. Doing this quickly and securely, at scale is a competitive advantage. The recent proliferation of Big Data technologies has ensured that banks can have the capability to process huge volumes of unstructured data, typically in the form of social media, e-mails, call center data, SMS messages, customer survey & feedbacks etc.
With an increasingly large number of people willing to adopt multi-channels like complaint logs, feedback forms, IM Chats, social media for conducting business, it has made a fundamental shift in the way banks interact with prospects, customers, employees and other stakeholders. These channels contains information in the form of unstructured text, images, and subscriptions to specific communities, and so on. It is very important for Banks to understand the techniques that can be applied to a given type of data to extract meaningful information and build Customer analytics and Risk Management models on top of this. Also insights and analysis from external data like social media help banks analyze, understand, and predict market movements to enhance business outcomes. According to research by McKinsey, banks with advanced capability of using analytical solutions have a 4% – 6% point lead in market share over banks who do not.
However, perfecting these analytical solutions is not easy and banks are struggling to make meaningful inferences from increasing volumes of data and are only using a small portion of this unstructured data to generate insights that enhance the customer experience.
Unstructured Data insights for better connect with customers
Banks should have an effective strategy & framework and analyze unstructured data not only for awareness or marketing purpose but also to find early warning signal, competitive analysis, emerging trends, risks and threats.
For better connect and customer experience, the key is to integrate customer service, marketing and other banking processes with data driven insights. To attain that, banks should have an analytical platform to ensure intelligent decision making and customer targeting by mining unstructured data from varied sources.
The list of areas where unstructured data can be used for banking is restricted only by one’s imagination, but here are some areas we have seen emerge as top use cases:
- Deeper understanding of Customer preferences & sentiment
- Satisfaction Index & Level of customer service
- Topic of Interests (Supervised Deep Dives)
- Early Warning Signals
- Recommendation Engine
- Context Based Analysis
- Customer Churn
Transform Unstructured Data into actionable insights
There is no question that there is a great interest in mining unstructured data, but it’s a vast space in terms of breadth and depth. Companies may look at some of the below areas to understand what insights can be generated from unstructured data:-
- Mining disparate data sources for themes: Using data connectors & automated ETL process, companies can extract data from multiple sources, including data stored in enterprise systems such as Microsoft Exchange & SharePoint, other enterprise applications, as well as data within cloud services such as Salesforce.com, and Google Drive. One could mine the data to identify trends/threats, or to find references to topics of interest, and common themes across these disparate data sources
- Stock performance vs. Social Sentiment: There’s a growing movement around using social media sentiment as a stock guide. Social sentiment can be used as a helpful predictor of stock price fluctuations and supplement traditional fundamental and technical analysis. With help of sentiment analysis, one can get insight into what end customers are saying about brand’s products and services and analyze how sentiment changes over time
- Early warning signal on Incident Data: Nowadays, people are spending excess time on social media sites like Twitter, Facebook and other Forums sharing & seeking information on banking services. Banks can analyze tens of thousands of customer-service related comments shared on Web forums, blogs and other social media sites which can be harnessed as an early-warning signal of adverse incidents resulting from payments/transfers, login/sign-in issues, service not available etc. This will enable Banks as an early warning signal to reduce the incidents before it cascades into a major problem. In a field where any wrong banking transactions have significant liability and consequences, “the idea is to get the initial signals earlier, allowing them to investigate leads sooner, and resolve the issue & not escalate it.”
- Call center Analytics: keeping a customer happy is more important than ever before, customers are using multiple channels like Twitter/Facebook or phone calls to vent their opinions, grievances etc. Voice is still supreme but social media & online communities have changed the name of the “call center” to the “contact center”. Speech analytics on voice and sentiment scoring on voice/social platforms allows a business to detect an unhappy customer during a call or a tweet, predict that they are likely to churn, and extend a retention offer to keep that customer. Because of the advent of new social channels, fresh analytics approaches & methods are required for enhancing customer experience and retention
- Intelligent Business Operations with Video Analytics: With the advent of advanced video analytics capability, organizations now can identify events, attributes or patterns of behavior through video analysis of monitored environments to create intelligent business operations. With video analytics, companies are taking situational awareness to another level and optimizing customer service and operations with valuable insights around object detection, classification, behavioral patterns, threats, compliance etc. from recorded & streaming video data
- Enter the technique of graph analysis: Zynga uses HP Vertica in a massive cluster for social graph analysis to discover relationship of their customers – “the players of their games” that enables them to improve the player experience and average revenue per user. Here is a detailed post on this “Uplevel Big Data Analytics with HP Vertica”. Vertica provides a highly scalable and fast engine for iterative graph algorithms. These type of use cases are not only relevant to the Zyngas and Facebooks of the world but to any company or organization that has a connected user community. From online communities, forums to user-to-user call logs, social networks are everywhere and finding relationships, influencers & customer behavior is the key to monetizing them
With a strong framework & technology in place, Bank will be able to propose their best banking product at the right time according to user needs & interests.
Analyzing unstructured data is going to be increasingly important for banks as customers demand high levels of service. Strong strategy & framework will help Banks better leverage technology, data and analytics, which in return will help retain and delight customers.
Banks can no longer live in denial and not use unstructured data for insights. They needs to realize that their competitors are proactively integrating data driven insights for decision making, thus threatening their own existence.
As the banking business becomes increasingly digitized, it has opened many doors never available before. The idea of utilizing unstructured and structured data for analysis is gaining great momentum. This will help Banks to get a holistic view of customers and build a real-time recommendation engine to predict Customers’ next moves. Insights from unstructured data like social media, chats & emails and structured data (transactional data, will help banks to better understand user needs & preferences and optimize customer engagement and services to a level which was not possible earlier.
With a few banks taking the lead, the direction is set for other banks to have robust strategy and framework to offer better customer services through gleaning unstructured data sooner rather than later.
It is just the tip of the ice berg and the possibilities are many indeed.…..
Senior Director of Analytics, HP Analytics Data Labs
Senior Director, Hewlett Packard Analytics Data Labs. Rags have about 20 years of strong operating and commercial leadership experience in global business services, including in analytics and technology. In the last 8 years, Rags has run large analytics operations for clients in consumer industries and in solutioning analytics deals in the UK/EMEA markets for CPG and Private Equity clients. His passion is to partner with CXOs on their strategic initiatives and adding value through combination of analytics, business process solutions and change management tools. He can be reached at firstname.lastname@example.org
Social Media Practice Leader, HP Analytics Data Labs
With an entrepreneurial mind-set, RK began his career in a start-up based social media company – TooStep.com (a Professional Social Networking Site). Over the last 7 years, he has worked in an advisory capacity specializing in creating social media framework & insights and was also responsible for building and developing the analytics engine, particularly in the areas of web, mobile & social media in the enterprise & consumer space. He has substantial work experience in Digital Media, Analytics, Lead Generation and Brand Building. His educational background includes an M.B.A. in (Marketing & IT) and Bachelor of Engineering in Information Technology. He can be reached at email@example.com
Try deep learning using MATLAB